Machine Learning Training Available v0 (“MALTA”) License Agreement
The dataset licence has been completed. It is purpose-built to preserve the Maltese language while safeguarding contributors: you retain full ownership of your voice and likeness, prohibited from voice cloning, while ensuring that all AI models developed with this data remain open-source and free to the public.
Summary of Key Terms
This is a human-readable summary of (and not a substitute for) the full MALTA License Agreement.
Before accepting the agreement, please review the key terms below:
Your Rights are Yours
You retain full ownership of your voice, likeness, and any copyrights. You only explicitly waive your privacy and "Right of Publicity" claims strictly to the extent necessary for the data to be used for the permitted purposes.
Open, Free, and Share-Alike
The dataset will remain freely available to the public. Anyone who redistributes or modifies the dataset must share it under this exact same MALTA license (no additional restrictions allowed).
No Voice Cloning or Biometrics
Your data cannot be used for biometric identification or voice cloning (mimicking your specific, identifiable voice). This does not prohibit the training of general, non-identifiable Text-to-Speech (TTS) models.
Limited Use Only
Your contributions can only be used for:
- Training or testing machine learning models
- Linguistics research
Strict Openness Requirements
Any machine learning model trained, fine-tuned, or developed using this data (regardless of how little of it is used) must:
- Have publicly available open weights
- Allow for commercial and non-commercial use of those weights and that model at no cost
Transparency in Testing
Any machine learning model evaluated or tested on this data must publish its results publicly. Testing only does not necessitate the release of model weights or the model itself.
Right to Withdraw
You may withdraw from the agreement, but this only applies to future distributions. Existing copies, derivative works, and models already trained cannot be revoked or deleted.
No Guarantees or Liability
The dataset is provided “as is” without warranties. Contributors are not liable for any claims arising from its use.
Strict Enforcement & Jurisdiction
Any breach of these terms immediately revokes the license. The agreement is governed by the laws of Malta.
Full License Agreement
The complete legal text of the MALTA v0 License Agreement.
This MALTA License Agreement Version 0 (“Agreement”) is entered into by and between the individual or entity or institution contributing voice data (“Contributor”) and the recipient(s) of such data (“Licensee”), collectively referred to as the “Parties.”
By acting as a Contributor you explicitly agree that you have understood the terms of this Agreement and shall be bound by it.
By downloading, accessing, copying, or using the Licensed Data, the Licensee explicitly agrees to be bound by the terms of this Agreement.
Should the Contributor disagree with this agreement or any amendments made thereof, the Contributor can withdraw from the Agreement and have their Licensed Data duly deleted from future distributions of the dataset. Such withdrawal only applies prospectively; existing copies, derivative works, and trained models are unaffected, and licenses previously granted cannot be revoked.
1. Definitions
The audio recordings of Contributor’s voice, associated metadata, and any textual transcriptions provided as part of the dataset.
The use of Licensed Data strictly for:
- Training or testing machine learning models; or
- Linguistics research.
Any and all use of Licensed Data for the purpose of voice cloning, biometric identification, or other uses outside the Licensed Purpose.
The process of using machine learning or other computational techniques to synthesize speech that intentionally mimics the unique vocal characteristics, timbre, or prosody of a specific, identifiable human Contributor. This does not include the training of general Text-to-Speech (TTS) or automatic speech recognition (ASR) models where the goal is to synthesize generic, non-identifiable voices or understand speech.
2. License Grant
The Licensed Data is made available for free and shall remain open for public access.
The Contributor retains all copyright, intellectual property rights, and proprietary rights in and to the Licensed Data. No rights other than those explicitly granted in this Agreement are transferred. However, the Contributor explicitly waives any claims or rights of privacy and publicity (including the "Right of Publicity") strictly to the extent necessary for the Licensee to exercise the rights granted for the Licensed Purpose.
Subject to the terms and conditions of this Agreement, the Contributor hereby grants to the Licensee a worldwide, royalty-free, non-exclusive, irrevocable license to reproduce, prepare derivative works of, publicly display, publicly perform, and distribute the Licensed Data, solely for the Licensed Purpose.
Licensee is strictly prohibited from:
- Using Licensed Data for any voice cloning applications;
- Using Licensed Data to create models that do not meet the open-access requirements set forth in in this Agreement;
- Using the voice, likeness, or any other personal attributes of the Contributor beyond the scope of the Licensed Purpose.
If the Licensee redistributes, modifies, or creates derivative works of the Licensed Data, such distribution must be governed by this exact MALTA License Agreement or a later version of it. The Licensee may not apply any additional or more restrictive terms to the Licensed Data.
3. Publication and Openness Requirements
Any machine learning model trained, fine-tuned, or otherwise developed using the Licensed Data (regardless of the proportion of Licensed Data used in the overall training mix) must:
- Have publicly available open weights; and
- Allow for commercial and non-commercial use at no cost.
Any machine learning model tested on the Licensed Data must publish its results in a publicly accessible format.
4. Warranties and Disclaimers
The Licensed Data is provided “as is” without any warranties, express or implied, including but not limited to warranties of merchantability or fitness for a particular purpose.
The Contributor shall not be liable for any claims, damages, or liabilities arising from the use of the Licensed Data.
5. Termination
If Licensee breaches any term of this Agreement, this license terminates immediately.
Sections 2.2, 2.4, 2.5, 3, and 4 shall survive termination.
6. Jurisdiction
This Agreement shall be governed by and construed in accordance with the laws of Malta.